What Theory Is,What Theory Was: Whence and Whither Theory

Abstract

This essay offers a critique of Gabriel Abend’s call for more action and less abstraction in social theory. The critique questions Abend’s insistence on pragmatic presentism—research that can be used tomorrow—and his implicit rejection of our time-honored commitment to construct clarity and rigor in theory. Abend’s insistent focus on the present ignores a long history of pragmatism in both social and management theory that has served our profession well. More importantly, his argument misinterprets pragmatism as a rejection of theory when, in fact, pragmatism suggests precisely the opposite.

When top aides first questioned his relationship with White House intern Monica Lewinsky, former US President Bill Clinton said there “is nothing going on between us.” Asked by a Grand Jury to explain his lie, Clinton famously quipped “it depends on what the meaning of is, is.” Clinton’s sophomoric attempt to distinguish between what “is” and what “was” became a key factor in his impeachment. It also reflects a common rhetorical strategy of parsing words to rationalize a lie.

I was reminded of Clinton’s sophistry when reading Gabriel Abend’s essay, book (Abend, 2023) and paper (Abend, 2008) about the problems of contemporary social theory. Abend believes that social scientists waste too much time making hair-splitting distinctions in our constructs and our theories. Our time is better spent, he suggests, in active engagement with the world, creating impactful knowledge that can be “used tomorrow.” In making this argument, Abend joins the chorus of those management scholars who question the value of theory and encourage more empirical pragmatism (see Suddaby, 2014a, 2014b for a summary review).

Much of what Abend asserts is reasonable. His call to action evokes the legacy of American pragmatism in social theory first described by William James (1907) and John Dewey (1908). Dewey, in particular, stressed the practical application of knowledge and the importance of experimentation and experience in learning. He believed that education should prepare individuals for active participation in democratic society and for solving practical problems in everyday life.

There are, however, serious flaws in Abend’s view of theory. Foremost, his argument is not new. Social science has a long history of oscillation between moments of high theory and corrective swings to pragmatic empiricism. Second, there are serious flaws in Abend’s view of pragmatism. Neither James nor Dewey believed that we should abandon theory or construct clarity in our pursuit of pragmatic knowledge. Rather, they insisted on the integration of philosophy and theory into pragmatic practices of knowledge creation. Third, both Dewey and James viewed theory and constructs as processes of knowledge development and not as mere properties of linguistic concepts. This is an important distinction, one that seems to have escaped Abend’s attention, as much of his criticism is devoted to the futility of trying to define constructs by their linguistic properties.

Each of these flaws draws from a temporal sleight of hand, used by both Clinton and Abend, that essentially creates a false distinction between past and present. Both share a strategy of describing the world the way it “is,” or could be, and in a way that encourages us to forget the world the way it “was.” Abend overlooks decades of debate about the value of theory, the need for construct clarity, and how pragmatism has often been a helpful corrective to strike the right balance between grand theory and dustbowl empiricism. In this essay I explain why such historical context matters and why the “was” of the past “is” important for the present and the future of organization theory.

There is/was Nothing so Practical as a Good Theory

Abend decries the lack of common sense in our use of words. We cannot advance our knowledge of institutions, he suggests, if we waste time trying to define institutions. Abend’s concern about the utility of construct clarity is part of a broader complaint about the futility of making universal claims to knowledge in our theories. Both are good points. However, they are not new and few contemporary management scholars would disagree with them. Like most social theory, we have had our moments of aspiration to grand theory. However, those days are gone.

In his influential article titled On sociological theories of the middle range, Merton (1949) suggested that we abandon our claims to universal knowledge and, instead, aspire to “middle range” theories that “are close enough to observed data to be incorporated in propositions that permit empirical testing” (Merton, 1949, p. 39). Merton suggested that middle-range theories arise inductively from a researcher’s actionable experience in the field, rather than deductively from a “single all embracing theory of social systems” even though the middle-range theory, once articulated, “might be consistent with one” (ibid, p. 39).

Merton’s (1949) aspiration lives on in many communities of scholarship in organizational theory. Many management and organizational scholars embrace grounded theory in their research that encourages a theorization that is more closely connected to an empirical context (Glaser & Strauss, 1966). Prominent scholars have spoken out against the overproduction of theories that never get empirically tested (Hambrick, 2007). Others (Davis and Marquis, 2005; Doh, 2015; Mair & Seelos, 2021) suggest that organizational theory ought to focus on causal mechanisms instead of promulgating more grand theory. Each of these arguments offer a reasonable compromise to Abend’s notion that we are wasting time with theory and should find ways to produce knowledge more efficiently.

However, we must not conflate such empiricism with knowledge or science. The myth that science advances by empirical workers relentlessly probing and testing theory—i.e., standing on the shoulders of giants—was questioned by a host of social theorists, particularly American historian Thomas Kuhn (1967), who observed that scientific paradigms, and their theories, often persist long after they lose their ability to predict. Post-modernists critically wounded the myth of universal knowledge claims in the social sciences, noting an emerging “crisis of representation” in which both words and theories became detached from the reality they claimed to reflect (Baudrillard, 1994). Foucault (2020) made the same point by using historical (archaeological) methods to show how different categories (discourses) of power and knowledge mutually constitute each other.

While organization theory was once obsessed with universal knowledge claims, as Abend implies with his dismissal of the need to search for a single best definition, we have evolved with respect to our understanding of the role of theory. We have abandoned the notion of a theory of everything. In contrast to our colleagues in the “hard” sciences, we fully understand that our theories must flexibly adapt to changes in social behavior. We are more aware now that, in contrast to our colleagues in the “hard” sciences, where objects of study typically do not change as a result of theories, in the realm of social science, social behavior often changes in direct response to theory. Consider Donald MacKenzie’s (2008) brilliant demonstration of how modern financial theories directly changed the behavior of investors and ultimately helped trigger the financial crisis of 2008. MacKenzie concludes that theories are not “cameras” that capture empirical reality but instead are “engines” that produce it. Others (e.g., Martí & Gond, 2018) reinforce this point.

What we see in this brief and somewhat selective history of theory in the social sciences is that much of what Abend states to characterize the current state of theory in organizational studies is inaccurate and overlooks a long history of trying to find the right balance between grand theory and dustbowl empiricism. The pendulum has moved in both directions over the decades, and what we see today is a compromising truce that, at least for the moment, captures a degree of balance between the two. What Abend claims “is” the state of theory in the social sciences actually reflects what “was” its state. We cannot understand the current state of theory in social science without contextualizing it in its developmental history.

The dangers of a discipline abandoning theory are well known. In his historical review of the professions, Andrew Abbott (1988) observes that theory enabled physics to retain its status as an elite and independent profession, while the lack of theory relegated engineering to a profession dominated by large corporate control. Abbott observed that engineers—consistent with Abend’s advice—abandoned theory to physicists in pursuit of pragmatic solutions that can be put into practice tomorrow. The pragmatism of engineers reduced their ability to generate new technologies and allowed physicists the intellectual agility to produce x-rays, lasers, and the atomic bomb. Theory was the basis for all of these innovations and has proven to be incredibly impactful. As Kurt Lewin (1943) observed, there is nothing so practical as a good theory.

Construct Clarity or Construct Clarification?

A subsidiary part of Abend’s argument against grand theory is that we waste too much time in a futile effort to define constructs. Most of our core concepts, he suggests, defy definition. What, he asks, is an organization or an institution? The properties of these categories of phenomena are too broad to be fully captured by a single definition. In this respect, the problem of construct clarity is similar to the problem of grand theory. Our empirical observations are too diverse and complex to justify a universal definition. Notably Abend does not offer any solutions to this problem, apart from skipping this stage and moving directly to the collection and analysis of data.

Again, much of what Abend asserts is reasonable. It is difficult to define what an organization or an institution is. A big part of the problem is that organizations and institutions are constantly changing. Sometimes organizations or institutions change because of the knowledge we produce. More typically, they change as part of the normal adaptive dynamic of any social entity to changes in its material or social environment. As a result, any effort to define an organization or an institution by listing its essential properties is doomed to fail. We need another way.

A terrific example of this problem arose in the sociology of the professions. An early attempt to define a profession (Greenwood, 1957) identified five distinguishing properties of any profession. Over time, however, new research added more defining properties to the list. Eventually the number of properties that defined a profession became so long that Wilensky (1964) suggested that the concept lacked coherence and asked if the research on professions simply demonstrated “The professionalization of everyone?” Wilensky’s clever title foreshadowed the answer to this problem and, as it turns out, the answer to Abend’s problem. Any effort to use static properties to define a rapidly evolving phenomenon is doomed to fail. A profession, like an organization or an institution, is an ongoing and evolving outcome of a dynamic process of change. When the objects of our study change rapidly, we must attend to processes by which they change, not their designated properties.

Andrew Abbott (1988) recognized this fundamental flaw in logic and his book The System of Professions almost immediately stopped the effort to define professions by a list of static properties. Professions change, Abbott observed, and instead of wasting time on applying static terms to dynamic flows, we should try and discern the historical processes that create professions. Karl Weick (1995) and Gareth Morgan (1997) provided similar insights with respect to the problematic definitions of organization. Verbs, not nouns, are better able to define things that are constantly adapting. Both advocate for the use of gerunds—organizing, professionalization—or infinitives—to organize, to professionalize—as the solution.

The logical contradiction that Abend creates for himself is failing to attend to the temporal tension between “is” and “was.” Some phenomena are highly amenable to statements of “is.” A geologist can easily declare that a sedimentary rock is a “type of rock formed by the accumulation or deposition of mineral or organic particles at Earth’s surface.” The definition holds across time (or, human time, at least) and space (well, earthly space at least). There is no need to clarify the temporal contingency of that definition. Social theorists, however, must be careful to define terms like gender or race with some sensitivity to both “is” and “was” as these constructs change meaning across both space (culture) and time. Any claim to define these constructs universally—e.g., with a statement of “gender is” or “race is” must be prepared for the Clinton-esque challenge that “it depends on what your definition of “is” is.

Many of these semantic challenges go away when we embrace a processual or historical view of our objects of study. Precision of definition becomes easier when we contextualize our constructs in space and time. Dewey (1908) likewise believed that knowledge is not static but rather dynamic and contingent upon its context. He argued that knowledge is continuously evolving through human experience and interaction with the environment. James (1890) was similarly interested in understanding how different cultural and temporal contexts influence human consciousness and behavior. He explored the diversity of religious experiences across cultures and highlighted the variability of human beliefs and values.

Because of their emphasis on contextualized knowledge and the role of learning as a dynamic experience, both Dewey and James emphasized the role of processes in creating knowledge. There is a substantial foundation for process theory and research both in organization studies (Langley, 1999; Pettigrew, 2012; Van de Ven, & Poole, 1995) and in social theory (Cloutier & Langley, 2020). Our profession could indeed still embrace processes more fully. Doing so would help us better incorporate notions of “is” and “was” into our theoretical frames. Clearly, however, we need not abandon construct clarity simply because we conflate “is” and “was.” We can help define what a particular institution “is” by adopting a processual analysis of what the institution “was.”

If Construct Clarity is so Bad, why is Abend so Good at it?

One of the curious aspects of Abend’s essay, book and paper is how he brilliantly uses construct clarity to explain why we do not need it. For someone arguing against construct clarity, at least in a universal sense, Abend spends a good deal of time parsing words. Abend is careful to define terms, draw boundary conditions around them, and describe their relationship to other terms, all outstanding examples of how to achieve construct clarity (Suddaby, 2010), and all in the service of explaining why construct clarity is a waste of time. This is odd and begs the question of why Abend feels the need to use the tool he is criticizing to make his point.

There are two possible explanations. The generous one is that Abend disavows the use of construct clarity but feels the need to embrace it because it is the lingua franca of his audience. The more likely one is that, despite Abend’s protests, clearly expressed constructs actually help social scientists talk to each other. Construct clarity is acutely important to social science but is not unique to it. Our colleagues in other disciplines have a much more advanced understanding of the value of conceptual abstractions of phenomena that we cannot directly observe, invented for a specific scientific purpose (Suddaby, 2010). Their value is that they impose meaning on the chaos of the empirical world. They take the “blooming buzzing confusion” (James, 1890, p. 462) of empirical reality and carve it into “robust categories that distil phenomena into sharp distinctions.”

Our colleagues in the hard sciences have employed construct clarity very effectively. Consider physicists who, based on painstakingly detailed calculations, come up with summary concepts like quarks or leptons and then construct the equipment required to prove that they exist. Some of their constructs are so universally powerful—consider hysteresis from magnetic theory as an example—that they work their way into the social sciences, e.g., psychology (Farrell, 1999), economics (Darity & Goldsmith, 1993), or sociology (Bourdieu, 1990), with often little loss of power or meaning.

Construct clarity is both simple and effective. Its intent is to pragmatically harness the nuances of everyday language, strip away its surplus meaning (MacCorquodale & Meehl, 1948), and strategically adapt it to allow scientists to understand each other. Yes, there is always a danger that we get into hairsplitting arguments about the number of angels that can dance on the head of a pin, or debates about what the meaning of “is” is. However, those philosophical debates are rare, particularly in organization theory, where the philosophical tendencies of some scholars typically become subordinate to the more practical art of taking abstract ideas and operationalizing them into granular entities that can be empirically observed and measured.

Whence Theory?

Some time ago, I wrote an editorial essay to address concerns very similar to those raised by Abend in his essay and book. The essay was titled “Why Theory” and raised many of the points presented in response to Abend’s call for a more pragmatic social science. Since then, advances in both Big Data and Artificial Intelligence have reinforced the relevance of, and need for, theory. Advances in AI and the growth of massive data sets now threaten the tradition of scientific method that underpins research methods and practice in most universities. Advocates of this new approach support Abend’s argument of the need to adopt a pragmatic approach to science that eschews theory and focuses on practical results that can be “used tomorrow.”

Perhaps the best example of this is the legitimacy contest that emerged between two models of science in the race to map the human genome. American geneticist Craig Venter became frustrated with the slow pace of progress made by the National Institute of Health in mapping gene sequences using traditional scientific methods (Anderson, 2000). Venter formed a private company and used a new technique of “shotgun sequencing” which applies statistical techniques to identify recurring patterns in large samples of randomly fragmented DNA segments. Computer programs use the overlapping ends of different fragments to assemble them into a continuous sequence. The technique is more efficient than traditional methods of sequencing because the process ignores the tedious work of reading theory, creating and then testing hypotheses. Instead, it applies massive amounts of analytic power to randomly identify correlations, a process we used to dismiss as “data mining.” Venter’s method produced results that could be “used tomorrow.”

Data mining now threatens to replace the scientific method (Smith, 2020). Analysts use these techniques for everything from predicting consumer behavior to predicting what words we will use next when composing an email. Some scientists celebrate this as the next big leap in science, while others worry that it will be its end because it sacrifices expedient prediction for true understanding (Cornelissen et al., 2024; Wise & Shaffer, 2015). Notably, data mining threatens the “hard” sciences much more than the social sciences. Still, it satisfies Abend’s appetite for science that can be used tomorrow.

Prior experience, however, suggests that we cannot produce knowledge for tomorrow without attending to the past. This is Andrew Abbott’s (2001) argument in Time Matters: On theory and method where he decries the dominant approach to research in sociology that assumes the world consists of static entities. Our mission as researchers, in the traditional worldview of positive social science, is to identify the meaning and causal purpose of the attributes independent of time and context. Abbott, however, argues that this assumptive view of the relationship between social science and reality is flawed. The sequence of events matters. The context of events matters. History matters.

Perhaps the biggest issue with Abend’s essay is that it ignores a long history on the subject of pragmatic theory in the social sciences. Sociology, in particular, has struggled with useful interpretations of pragmatism in theory. Mills (1959) spoke of abstracted pragmatism to emphasize the need for sociologists to strike a balance between theoretical frames and concrete empirical reality, while remaining sensitive to issues of power. James spoke of radical pragmatism to capture pluralism in theory, the idea that there are multiple valid interpretations of reality. Abend’s version of pragmatism is not clear but seems to most closely resemble the pragmatism of Craig Venter. This might work in the hard sciences, but I question its utility in the social sciences. I worry that, beyond mere utility, the Abend–Venter form of pragmatism may be dangerous. Venter’s pragmatism works best in generating objective knowledge, which seeks knowledge of the world as it is. However, some disciplines hope to generate normative knowledge, which seeks knowledge of the world as it might be. Organizational theory in my view ought to generate syncretic knowledge, which creatively combines objective and normative worldviews.

More generally, I worry about science that sacrifices predictability for understanding. My worry comes back to the difference between “is” and “was.” AI’s efficiency and accuracy depends on data collected in the past. That is, it depends on what “was” and it will work best in stable contexts—i.e., in knowledge domains that do not change quickly. It will not work well in studying dynamic contexts and may not work at all in contexts that are self-aware—e.g., in communities of people who know their behavior is under study and react to it. We should never underestimate human ingenuity in countermanding totalitarianism, or purposely trying to confound governance by algorithm. The answer to resisting this new world requires a high degree of sensitivity to the relationship between “is” and “was.” Ultimately, our ability to overcome the violence of a hyper-pragmatic science may rest on our ability to manipulate how algorithms interpret our past behavior. That is, it may well depend on what the meaning of what “was” is.